Chaos in anthropology.

Mike Diehl (V027JW4C@UBVMS.BITNET)
Mon, 11 Apr 1994 16:02:25 -0500

I had intended a longer post, but UB's !*excellent* network server
dumped me on my virtual can about three lines from the completion of my post.
So, in *brief*:

1) Background (mine): 10 months at SFI as a Doctoral Fellow. I'm a SW
2) IMHOp there exists no good example of Chaotic behavior in any
anthropologically or archaeologically documented system. This is a
consequence of one or two things that I will mention below. Also, I
am *extremely* leary, wary, and skeptical of all uses of the word
CHAOS by all anthropologists, and more generally by almost everyone,
and fear that we all must do alot more reading on CHAOTIC behavior
and on our respective subjects, lest we shame ourselves by jumping
prematurely on the CHAOS bandwagon. (Suddenly I find Danny Yee's
recent missives highly relevent).
3) The distribution of realms of relatively convincing cases of behavior
that *looks* chaotic is small. Chiefly limited to: Agent-based computer
simulations that play an infinitely-iterated version of the prisoner's
dilemma game, with a built in capacity for agents to "learn." (See
Mats Nordahl's work at SFI). Pretty Mendelbrot sets generated by
software that interprets some highly specified and formally defined
mathematical systems that involve the mathematics of imaginary numbers,
and the study of fractal geometry. Some systems of fluid dynamics.
4) To demonstrate that a sytem behaves chaotically you have to have
*alot* of data in time series. Virtually all anthropological data
is of a synchronic nature, thus the dyanimics of the observed cultural
system cannot be discussed in any well-informed way. In a few cases, there
are intervals of a few hundred years for which someone (e.g. Dolores
Archaeological Program) has a reasonable amount of temporal information
(lots of dendro v and r dates) and a reasonably representative,
systematically and probabilistically achieved sample of data. To
my knowledge, no such data base has been demonstrated to have anything
resembling chaotic behavior. (No period doubling bifurcations etc.)

(I will continue this post in the subsequent messege).